Generation and analysis of novel Reln‐deleted mouse model corresponding to exonic Reln deletion in schizophrenia

Aim A Japanese individual with schizophrenia harboring a novel exonic deletion in RELN was recently identified by genome‐wide copy‐number variation analysis. Thus, the present study aimed to generate and analyze a model mouse to clarify whether Reln deficiency is associated with the pathogenesis of schizophrenia. Methods A mouse line with a novel RELN exonic deletion (Reln‐del) was established using the CRISPR/Cas9 method to elucidate the underlying molecular mechanism. Subsequently, general behavioral tests and histopathological examinations of the model mice were conducted and phenotypic analysis of the cerebellar granule cell migration was performed. Results The phenotype of homozygous Reln‐del mice was similar to that of reeler mice with cerebellar atrophy, dysplasia of the cerebral layers, and abrogated protein levels of cerebral reelin. The expression of reelin in heterozygous Reln‐del mice was approximately half of that in wild‐type mice. Conversely, behavioral analyses in heterozygous Reln‐del mice without cerebellar atrophy or dysplasia showed abnormal social novelty in the three‐chamber social interaction test. In vitro reaggregation formation and neuronal migration were severely altered in the cerebellar cultures of homozygous Reln‐del mice. Conclusion The present results in novel Reln‐del mice modeled after our patient with a novel exonic deletion in RELN are expected to contribute to the development of reelin‐based therapies for schizophrenia.

relative amount of reelin in serum was lower in this subject 17 and established induced pluripotent stem cells (iPSC) from the subject. [25][26][27] Several clinical studies suggest that reduced reelin in the brain and peripheral blood might be associated with mental disorders. [28][29][30][31] Behavioral changes in Reln-mutant mice, reported by numerous studies, exhibit schizophrenia-like phenotypes (Table 1) 5,[31][32][33][34] ; however, insights into the biological mechanisms during brain development due to reelin reduction in humans remain limited. One clue regarding this issue comes from our recent quantitative single-cell time-lapse analysis of neuronal differentiation and migration using Reln-del iPSC revealing the developmental significance of reelin in migrating directionality. 25 To examine directional abnormality in neuronal migration observed in human RELN-deleted iPSC in vivo, we generated a mouse line in the C57BL/6J strain harboring the specific Reln deletion identified by us in a Japanese subject with schizophrenia, and named the line Reln-del mice. Our behavioral analyses revealed abnormal sociability in heterozygous Reln-del mice. This novel Reln-del mouse model may be valuable for understanding reelin signaling, and for studies on the pathological mechanisms of schizophrenia via the reelin deletion.

Animal experiments
All research and animal care procedures were approved by the Nagoya University Animal Care and Use Committee. Mice were housed in groups of maximum six animals per cage and maintained on a regular 12-h light/dark cycle (09:00 to 21:00 hours light period) at a constant 23 C temperature. Food and water were available ad libitum. Heterozygous Reln-del mice were generated by intercrossing Reln-del males and C57BL/6J females. Littermate WT C57BL/6J mice were used as controls, and 10-19-week-old Reln-del and WT mice were used in all experiments. All animal protocols were approved by the Animal Care and Use Committee of Nagoya University Graduate School of Medicine; in addition, the Principles for the Care and Use of Laboratory Animals, which were approved by the Japanese Pharmacological Society, and the National Institutes of Health Guide for the Care and Use of Laboratory Animals were followed.
Other materials and methods are described in Supplementary materials.

Results
Generation of Reln-del mice We previously identified a schizophrenia subject harboring a partial RELN deletion in chr 7 (102919640-102930809, NCBI36/hg18), which includes the region encompassing exons 52-58 encoding reelin repeats 7-8 (Fig. 1a). 17,24 We generated a genetically modified mouse corresponding to this deletion in RELN by genome editing using the CRISPR/Cas9 method; the strain was C57BL/6J, the standard murine line for general behavioral analyses. The Reln gene structure between human and mouse is highly conserved. First, we designed 2 CRISPRtargeted guide RNAs and a donor DNA for inserting a FLAG-tag sequence and a stop codon in exon 52 to generate the Reln-del mouse model (Fig. 1b,c). When two guide RNAs were transiently transfected into the mouse neuroblastoma cell line Neuro2a, we found that the #4 guide RNA showed the strongest cleavage activity by the T7E1 assay (Fig. 1d). We examined the possibility of the off-target effect of the #4 guide RNA using the CRISPR Design tool supported by MIT (http://crispr.mit.edu/) and confirmed that there were no pseudosequences due to a mismatch of less than three base pairs in the whole mouse coding regions (Table S1). The #4 guide RNA was chemically synthesized as a Crispr-RNA, and tracr-RNA, Cas9 protein, and the donor DNA were co-microinjected into 171 fertilized eggs in C57BL/6J mice bred at Charles River, which were cultured up to the two-cell stage. Next, we transplanted 119 eggs into the oviducts of pseudopregnant mice, and 18 littermates were born. As a result, we confirmed that six knock-in mice were obtained as designed by polymerase chain reaction (PCR) cloning (Fig. 1e).
Validation of Reln-del mice Screening of the generated mice by conventional genotyping was achieved with three primers designed to distinguish wild-type (WT) and mutant (Reln-del) alleles by agarose gel electrophoresis ( Fig. 2a,b, Table S2). Homozygous Reln-del mice were born normally and exhibited normal growth until 21 days after birth. However, Relndel mice exhibited phenotypic similarities to staggerer and other reeler mice (Fig. 2c, Movie S1). We also validated the reelin protein expression level in the novel Reln-del mice by immunoblotting (Fig. 2d). In embryonic day 17 and postnatal day 21, the expression of reelin in heterozygous Reln-del mice was approximately half the reelin expression in WT mice; reelin was detected barely in homozygous Reln-del mice in those ages. Expression of vinculin as an internal control was comparable between the genotypes. The specific reelin antibody used in the current study could not detect lower molecular-weight bands by immunoblotting in homozygous Reln-del mice, indicating that they were null mutants with no truncated reelin.
Histological analysis of Reln-del mice We investigated histological changes in Reln-del mice. Hematoxylin/ eosin-stained brain slices of homozygous Reln-del mice showed severe brain malformation (Fig. 3). Specifically, cerebellar atrophy, enlargement of the cerebral ventricles ( Fig. 3a-c) and cerebral dysplasia ( Fig. 3d-f), which were one of the typical reeler phenotypes [38][39][40][41] were remarkable. In addition, we observed hippocampal malformation and disruption of the dentate gyrus and granule layer ( Fig. 3g-i). On the other hand, heterozygous Reln-del mice did not show severe brain malformation like homozygous Reln-del mice. Behavioral analyses of Reln-del mice Reln mutation is a risk factor in several mental disorders, such as schizophrenia and autism spectrum disorder. Therefore, we performed global behavioral analyses using the following 10 behavioral tests to determine whether Reln-del mice were a relevant model for schizophrenia: locomotor, rotarod, open field, elevated plus maze (EPM), social interaction, marble burying, PPI, Y-maze, novel object recognition, and fear conditioning tests.
We used heterozygous Reln-del mice for all behavioral analyses, because nearly all homozygous Reln-del mice were hypoplastic and had difficulty in reaching the adult stage. Additionally, the original Japanese subject that was the basis of our studies harbored a heterozygous Reln deletion. 17 First, we evaluated the general locomotor function in heterozygous Reln-del mice, because homozygous mice showed severe incoordination and cerebellar atrophy, which may change motor, emotional, and cognitive function. Heterozygous Reln-del mice showed normal locomotor activity and coordinated movement in the locomotor and the rotarod tests ( Fig. 4a-d). These data show that we could evaluate emotional and cognitive functions of heterozygous Reln-del mice with normal motor function.
A previous study has reported that heterozygous reeler mice have reduced anxiety. 42 Therefore, we evaluated anxiety in the open field and EPM tests. Typically, WT mice spent more time in the periphery than the center of the circle in the open field test and preferred the closed arms compared with the open arms in the EPM test, given that the center of the circle and the open arm were sources of anxiety in mice. Importantly, there were no significant differences between the WT and Reln-del mice in both tests (Fig. 4e-i). To investigate anxiety and irrational behavior, we performed the marbleburying test and found that Reln-del mice showed normal behavior (Fig. 4j).
Impairments of social behavior and sensorimotor gating function are the main symptoms of schizophrenia. To examine these points of heterozygous Reln-del mice, we performed the three-chambered social interaction and the PPI tests. There were no significant differences in habituation or sociability sessions between the WT and heterozygous Reln-del mice (Fig. 4k,l). In the social novelty session, WT mice significantly preferred the novel mouse (stranger 2) than the familiar mouse (stranger 1). Conversely, heterozygous Reln-del mice showed the same interest in both strangers 1 and 2 (Fig. 4m). In the PPI test, heterozygous Reln-del mice exhibited a phenotype similar to that of WT mice (Fig. 4n,o). The difference may have caused low attention to stranger mice, but not anxiety to the strangers and some kind of recognition, as the levels of WT and heterozygous Reln-del mice were not different in open field ( Fig. 4e-g), EPM (Fig. 4h,i), or novel object recognition tests (Fig. 4p,q). Further in heterozygous Reln-del mice, Reln-del mice exhibited no significant phenotypes in memory function in the Y-maze and fear-conditioning test (Fig. 4r-v).

Cerebellar granule cell migration in Reln-del mice
We performed the time-lapse analysis to assess migration in cerebellar granule cells isolated from Reln-del mice and conducted biochemical analysis of reelin signaling as well. Reelin binding to its target receptor, such as apolipoprotein E receptor 2 and very-low-densitylipoprotein receptor (VLDLR), leads to the phosphorylation of the adaptor protein Dab1. 43 We prepared cerebellar granule neurons from postnatal WT and Reln-del mice for reaggregation, which is often used to study neuronal migration, 44 and conducted a 3-day time-lapse analysis (Fig. 5, Movies S2-S4). Then, we performed automatic recognition and quantification of the cell body and neurite length (Fig. S1) (Fig. S1); however, reaggregates from homozygous Reln-del mice were affected by cerebellar hypoplasia and showed severe structural defects compared with those from both WT and heterozygous Relndel mice; therefore, homozygous Reln-del mice were difficult to quantify (Fig. 5c, Movie S4). Next, we analyzed the reelin signaling pathway in these cerebellar neuronal reaggregation cultures. First, we confirmed that reelin was highly expressed in WT cultures and that   (c) Appearance of Reln-del mouse 21 days after birth. Homozygous Reln-del mice staggered (see also Supplementary Movie S1). (d) Immunoblot analysis to confirm reelin deletion in Reln-del mouse brain. Fifty-microgram total lysate from the whole brain of postnatal day 21 mice were loaded in each well. The signal was detected by an Odyssey imaging platform. The molecular weight of full-length reelin protein is approximately 388 kDa, which is indicated by the green signal. The internal control vinculin (117kDa) is indicated by the red signal. The expression level of full-length reelin relative to that of vinculin was calculated by the software attached to Odyssey and is shown in yellow font.

PCN Psychiatry and Clinical Neurosciences
Phenotypes of novel Reln-deleted mouse the reelin protein expression level was reduced in cultures from heterozygous Reln-del mice (Fig. 5d), whereas the expression levels of the internal control vinculin and the dopaminergic neuronal marker tyrosine hydroxylase were unchanged (Fig. S2a). Finally, we examined the reelin signaling quantitatively by analyzing the relative expression levels of p-Dab1 and total Dab1 (Figs. S2b,c) and found that the phosphorylation of Dab1 was likely impaired in cultures from heterozygous Reln-del mice.

Discussion
Disorders of neuronal migration in humans result from the disruption of the normal movement of neurons from their original site of birth to their final destination during early development. 45 Mutations of many genes are involved in neuronal migration disorders and include LIS1 and DCX in classical lissencephaly spectrum, TUBA1A in microlissencephaly with agenesis of the corpus callosum, and RELN and VLDLR in lissencephaly with cerebellar hypoplasia. 8,45 Genetic studies on rare variants implicate that RELN is associated with schizophrenia. 19,46,47 Furthermore, our copy-number variation analysis identified a subject with schizophrenia who harbored a novel exonic deletion in RELN. 24 As reelin plays important roles in various pathways from development to postnatal function, Reln-del mice are expected to be a reasonable schizophrenia model that can reveal the pathogenic mechanisms underlying schizophrenia. Future studies can facilitate accumulation of sufficient genetic data and clinical phenotypes in humans for elucidation of the molecular mechanisms of schizophrenia caused by a decrease in reelin expression.
In the current study, we generated a novel Reln-del mouse model to mimic the Japanese subject with schizophrenia harboring the specific exonic deletion in RELN and compared it with the findings of previous studies on other reeler-associated models, including Orleans reeler mice (Table 1). Specifically, we designed a Reln-del mutant allele carrying mutations to code a FLAG-tag and stop codon in exon 52, corresponding to the amino acid residues within the seventh and eighth reelin repeats (Fig. 1). The homozygous Reln-del mice exhibited the typical phenotypes of the staggerer as well as the classical reeler mice (Fig. 2c). 38 Immunoblotting of whole-brain lysates using an anti-reelin antibody recognizing the N-terminal region of reelin revealed that the endogenous reelin level, which was barely detectable in homozygous Reln-del brain lysates, was approximately half of that in WT mice in heterozygous Reln-del mice (Fig. 2d). In contrast, Orleans reeler mice have a transcriptional reelin product with a 220-bp deletion within the seventh and eighth reelin repeats and potentially express a truncated reelin. 11,12 Our novel Reln-del mice more faithfully reflect the Japanese subject with the RELN deletion compared with our previous study using Orleans reeler mice. 17 In addition, we generated the Reln-del mice in the C57BL/6J strain that is commonly used for behavior studies in schizophrenia models, distinct from Orleans reeler mice, which were maintained in the  Balb/c strain. Differences of inbred strains and/or reelin products might underlie differences in the behavioral phenotypes between the mouse models (Table 1). Histological analysis showed cerebellar atrophy, hippocampal dysgenesis, and cerebral dysplasia in homozygous Reln-del (Fig. 3). These phenotypes are similar to those of both Jackson and Orleans reeler homozygous mice. [38][39][40][41] On the other hand, we found that heterozygous Reln-del mice did not show severe brain malformation like homozygous Reln-del mice (Fig. 3). These results are similar to those reported by previous other studies. 38,41 As the results of global behavioral analyses, we found abnormal social novelty in heterozygous Reln-del mice (Fig. 4), indicating that heterozygous Reln-del mice partially mimicked the Japanese subject with schizophrenia who harbors the same RELN mutation. Several studies reported that Relnrl-J/+ mice showed abnormal behaviors, such as PPI deficits, 5,48 unusual anxiety, 42 and reduced cognition and sociability. 5,16 These differences may be associated with the novel RELN mutation in heterozygous Reln-del mice. However, the expression pattern of reelin protein in heterozygous Reln-del was similar to that of Relnrl-J mice based on our immunoblot analysis, which indicate that the reelin-deficient phenotype might be minimally affected by the growth environment or genetic background. It was reported that reelin related with neuronal plasticity and spine formation. 7 Most heterozygous reelin mutant mice, including our mice, showed  Fig. 4f,i; n = 12 for WT mice, n = 13 for Reln-del mice in Fig. 4j; n = 24 for WT mice, n = 26 for Reln-del mice in Fig. 4n,o; n = 12 for WT mice, n = 13 for Reln-del mice in Fig. 4p,q; n = 22 for WT mice, n = 22 for Reln-del mice in Fig. 4r,s; n = 12 for WT mice, n = 13 for Reln-del mice in Fig. 4t-v). (k-m) Performance in the three-chambered social interaction test: (k) habituation, (l) sociability, and (m) social novelty sessions. Data represent the mean AE SEM (n = 18-20 for WT mice, n = 20 for Reln-del mice). ***P < 0.001 between empty chamber and stranger 1 or *P < 0.05 between stranger 1 and stranger 2 (two-way analysis of variance with post-hoc Sidak's multiple comparisons test).

Clinical Neurosciences
Phenotypes of novel Reln-deleted mouse abnormal behaviors without clear histological changes. 5,17,31,33,34 These kinds of abnormal behaviors may be caused by minimal changes, such as neuronal plasticity and spine formation. Approximately half the levels of reelin may be sufficient for normal neuronal development in the critical developmental stages because reelin signaling includes autoinhibition systems, such as Dab1 reduction. However, reelin is invaluable in the adult stage and has some neuronal function because expression levels are less than those observed in the developmental stage and distribution is different. These phenomena suggest that there is a lack of gene-dose dependency in mice with Reln-del.
A study in humans showed that patients with lissencephaly carrying RELN mutations in splicing recognition sites also exhibited cerebellar hypoplasia due to a decrease in full-length reelin, 8 in agreement with the features such as altered neuronal differentiation and migration observed in our in vitro phenotypes. 44 Specifically, we showed severe structural defects in homozygous Reln-del cultures ( Fig. 5c and Movie S4), which closely associate with cerebellar hypoplasia, 8 indicating a potentially reliable lissencephaly model. Conversely, neurons in heterozygous Reln-del cultures are expected to be more likely a good model of schizophrenia based on genetic studies. 17,24 Neurons in heterozygous Reln-del cultures showed an unmarked phenotype of migration abnormality, but there were no differences in gene dose between neurite migration and elongation in vitro (Fig. 5a,b, Fig. S1, and Movies S2,S3).
The iPSC technology is a promising tool for creating a new generation of pathophysiological assays for in vitro drug screening, which allows time-course analysis of neurons from patient-derived iPSC that are not readily available in human studies, and recent studies utilizing iPSC facilitate the interrogation of molecular pathways and cellular dynamics in humans. 49 Nonetheless, the pathogenesis of schizophrenia remains poorly understood, and analysis in animal models is necessary for understanding the neural circuits and tissue structures in schizophrenia. Using reliable and predictive mouse models is essential to understand the neurobiological basis of schizophrenia and the development of novel drugs with improved therapeutic efficacy. 50 Our previous trajectory analysis of the highly homologous neurons, such as tyrosine-hydroxylase-positive neurons from iPSC, is one of the best ways to understand the mechanisms of microstructural defects in the brains of subjects with schizophrenia. 25,51 Cerebellar neurons from homozygous Reln-del mice showed a very severe migration defect that was similar to the outgrowth abnormality of tyrosine-hydroxylase-positive neurons derived from iPSC of the Japanese subject with Reln-del. In our in vitro analysis, the cell bodies and neurites of the cerebellar reaggregates were too fine, and the population was too dense, which hindered the measurement of the area of each cell body and the trajectory analysis. Overcoming these technical issues will allow for more precise comparisons of neuronal dynamics and morphology in these cultures with those of human iPSC-derived neurons. In vitro analyses alone are insufficient to elucidate the molecular pathology of individuals by behavioral analyses of schizophrenia model mice. Advantages of highly homologous differentiation of human iPSC into specific neuronal subtypes will significantly aid in consideration of in vitro analyses, such as reelin signaling during migrating, probably promoting the development of reelin-related therapies.
In conclusion, our in vitro and in vivo data indicate that the novel Reln-del mice, using CRISPR/Cas9, which partially mimicked schizophrenia-like behavior, provide a novel platform for elucidation of the pathogenesis underlying schizophrenia and the development of reelin-based therapies.

Supporting information
Additional Supporting Information may be found in the online version of this article at the publisher's web-site: Table S1. Mutations in coding regions at putative off-target loci containing up to 4 bp mismatches for Reln target are listed. This list was calculated by following CRISPR design tool: http://crispr.mit.edu/. Mismatches compared to on-target sequence are shown in red. PAM sequences are labeled in green.